Human Reproduction Update Advance Access originally published online on June 28, 2006
Human Reproduction Update 2006 12(5):499-512; doi:10.1093/humupd/dml027
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Screening in women’s health, with emphasis on fetal Down’s syndrome, breast cancer and osteoporosis
1 Department of Obstetrics and Gynecology, Erasmus University Medical Center, Rotterdam, The Netherlands, 2 Academic Unit of Primary Health Care, Department of Community Based Medicine, University of Bristol, Clifton, Bristol, UK and 3 Department of Obstetrics and Gynecology, University of Kansas School of Medicine, Kansas City, KS, USA
4 To whom correspondence should be addressed at: Department of Obstetrics and Gynecology, Erasmus University Medical Center, PO box 2040, 3000 CA Rotterdam, The Netherlands. E-mail: h.wildschut{at}erasmusmc.nl
 |
Abstract |
|---|
 TOP Abstract IntroductionPrenatal screeningBreast cancerOsteoporosisConclusionsReferences |
|---|
Screening tests have become increasingly popular in women’s health care over the last two decades. The initiative for screening is typically generated by either an agency or the health care professional being consulted for some reason. In many instances, however, the demand for screening tests is patient driven with the health care provider being poorly prepared to determine the usefulness of screening. This review illustrates the complexity of screening using three disorders where early detection and treatment have the potential to improve the quality and longevity of life. Prenatal diagnosis of Down’s syndrome does not offer the parents the opportunity for cure but does offer the opportunity for education and rational choice as the impact of the diagnosis on the family is weighed. The evidence for breast cancer screening is more persuasive for older than younger women, but even in older women, there is a balance of risks and benefits. Treatment options for osteoporosis have improved in terms of reductions in fracture risk as well as beneficial effects on bone density, but evidence of the effectiveness of a screening programme for this condition in an unselected population is lacking. Ultimately, it is crucial that women be provided with clear and comprehensive information about the screening programme, in terms of possible gains but also costs of various kinds: physical, economic and psychological.
Key words: breast cancer / Down’s syndrome / informed choice / osteoporosis / screening
|
Introduction |
|---|
|
TOPAbstractIntroductionPrenatal screeningBreast cancerOsteoporosisConclusionsReferences |
|---|
Screening tests have become increasingly incorporated into women’s health care over the last two decades. The growth has been of particular note in obstetrics with the introduction of screening tests for Down’s syndrome, fetal malformations and preterm birth and in gynaecology with large-scale screening for osteoporosis and breast cancer. The concept of screening is different from diagnosis. Although diagnostic tests are applied to patients who actively seek health care services to identify the cause of their illness, screening tests focus on individuals with no known and/or reported symptoms or complaints related to the condition of interest.
Screening is defined as a procedure to help identify, in an organized way, a specified disease or condition among asymptomatic individuals (Peters et al., 2006
). In general, the screening process is initiated by either an agency or the health care professional being consulted for some reason. Nonetheless, in many instances, the demand for these tests is patient driven with the health care provider being poorly prepared to determine the usefulness of screening from a societal perspective. Moreover, the individuals screened are not usually familiar with the disease for which they are being screened.
Apart from its potential virtues, screening is also associated with several problems including the generation of false-negative test results and undue anxiety secondary to false-positive test results (Green et al., 2004). Thus, one of the prerequisites of a successful screening programme is the provision of objective and balanced information on its potential benefits and limitations. Such information should include a clear and well-defined statement of the condition of interest, the nature, validity and reliability of the screening test(s), the implications of both normal and abnormal test results, the effectiveness of early treatment and the tangible and intangible costs. In practice, however, such information is often ambiguous or simply lacking. Moreover, these issues are often considered purely from the perspective of public health with little or no attention paid to the influence of individuals’ valuations or utilities relating to aspects of the screening process and potential outcomes. For example, the balance of risks and benefits from mammography should take into account the views of the individual woman, rather than emphasizing the need for full coverage as an end in itself. This does not detract from the need for good organization in any successful screening programme. Indeed, it emphasizes this prerequisite, because the provision of comprehensive and balanced information depends on good organization just as much as do the procedures for testing and subsequent management.
Returning to more general aspects of screening, it remains essential that certain standard criteria be met for a screening programme to be worthwhile (Wilson and Jungner, 1968). The first criterion is that it should address an important health problem in terms of occurrence and health implications, whereas the second is that there should be an acceptable and adequately successful test to distinguish those with and without the condition of interest. Moreover, there should be facilities and treatment options available to influence the future course of the condition over and above that which would be achieved without the screening programme. Finally, the programme should be cost-effective (Peters et al., 2006).
There are several reasons for varying conclusions on the effectiveness of screening programmes. These include (i) variations, including wrong estimates, of the prevalence of the condition of interest, which impacts the number needing to be screened to detect the event of interest; (ii) variations in uptake or acceptability of the screening tests; (iii) biased results from poorly designed and reported studies; (iv) lack of adequate facilities for confirming the diagnosis and or for adequate treatment and (v) lack of adequate clinical follow-up. Other difficulties in the interpretation of a test’s performance relate to the choice of an appropriate reference standard. The reference standard (gold standard) is considered to be the best available method for establishing the presence or absence of the condition of interest (Bossuyt et al., 2003). Verification bias often leads to overestimates of test performance, where those tested are more likely to be investigated further and hence have the diagnosis confirmed compared with those not tested. For example, women whose mammography results are deemed ‘normal’ are less likely to be subjected to an additional diagnostic workup compared with those with ‘abnormal’ results; the resulting underestimate of false negatives would artificially improve the performance of the screening test.
This review is focused on the essentials of screening for three important topics in women’s health: Down’s syndrome and other chromosomal abnormalities, breast cancer and osteoporosis. To be consistent with the second edition of our recently published text ‘When to Screen in Obstetrics and Gynecology’ (Elsevier Ltd, ISBN 10: 1-4160-0300-2), we present the available information for each condition in a standardized format, addressing the fundamental questions that both the practitioner and individual should consider to make an informed decision. For each topic, the literature search included relevant Cochrane databases and PubMed with no date restriction, using search terms relevant to the specific screening programme and section of the standardized format [e.g. incidence/prevalence, diagnosis and management, randomized controlled trials (of screening and of treatment) and cost-effectiveness]. Related articles were then sought from the most relevant publications.
|
Prenatal screening |
|---|
|
TOPAbstractIntroductionPrenatal screeningBreast cancerOsteoporosisConclusionsReferences |
|---|
Condition of interest
Trisomy 21, or Down’s syndrome, is the most common chromosomal abnormality among live born infants. The birth prevalence of Down’s syndrome depends on the maternal age distribution of the population being considered, as the probability of trisomy 21 increases with the age of the pregnant woman (Table I) (Morris et al., 2002; Grijseels et al., 2004). For instance, the proportion of pregnant women aged 
36 years is 14% in the Netherlands. As a result, the birth prevalence of Down’s syndrome among the Dutch population is higher than that expected among populations whose relative frequency of older pregnant women is lower.
|
The birth prevalence of live born infants with Down’s syndrome also depends on the prenatal screening policy that is offered to the population and the readiness of women to have a pregnancy termination if their pregnancy is affected (Cheffins et al., 2000; Bell et al., 2003; Khoshnood et al., 2004; Siffel et al., 2004). The historical birth prevalence, which excludes the potential effect of pregnancy termination for Down’s syndrome, is 1 in 800 (Spencer, 2006).
The chances of Down’s syndrome are inversely related to the length of gestation because fetuses with Down’s syndrome are more likely to die in utero than non-affected fetuses. In fact, the probability that the fetus with Down’s syndrome will die in utero when diagnosed at the time of chorionic villus sampling conducted in the first trimester of pregnancy is 43% [95% confidence interval (CI): 31–54%], whereas the probability of fetal death is 23% (95% CI: 19–28%) when Down’s syndrome is diagnosed at the time of amniocentesis, which is conducted in the second trimester of pregnancy (Morris et al., 1999). This fetal attrition (Stein et al., 1986) must be taken into account when assessing the performances of the various screening programmes for the prenatal detection of Down’s syndrome (Spencer, 2001; Alfirevic et al., 2003).
Down’s syndrome is clinically characterized by a typical facies (i.e. epicanthal folds, flat nasal bridge, protruding tongue and open mouth), single palmar crease, hypotonia and mental retardation (Saenz, 1999; Roizen and Patterson, 2003.) Cognitive impairment varies notably (though the preponderance will be considered mentally retarded) and cannot be predicted at birth (Saenz, 1999). The diagnosis of Down’s syndrome is confirmed by chromosome analysis. The exact cause of non-disjunction that leads to Down’s syndrome remains unknown. About half of children with Down’s syndrome are born with congenital heart disease, the most common being atrioventricular septal defect (45%), followed by ventricular septal defect (35%), isolated secundum atrial septal defects (8%), isolated persistent patent ductus arteriosus (7%) and isolated tetralogy of Fallot (4%) (Freeman et al., 1998). Other common features of Down’s syndrome include duodenal atresia and non-immune hydrops. Infants and children with Down’s syndrome are more likely to have hearing loss, otitis media, thyroid disease, ophthalmological disorders including congenital cataracts, skin disorders including palmoplantar hyperkeratosis, seborrhoeic dermatitis, fissured tongue, polycythaemia, transient myelodysplasia, acute myeloid leukaemia and acute lymphoblastic leukaemia, dental problems, coeliac disease, obesity, feeding difficulties, particularly if preterm or with concurrent cardiac and alimentary anomalies, neurological disorders including seizures and orthopaedic problems including osteoarthritic degeneration of the spine (van Allen et al., 1999; Saenz, 1999; Roizen and Patterson, 2003). About 8–13% of newborn infants with Down’s syndrome die in the first year of life (Julian-Reynier et al., 1995; Bell et al., 2003). Compared with Down’s syndrome infants without an additional anomaly, survival at 1 year is worse for those with additional anomalies (Bell et al., 2003; Wessels et al., 2003). During the last few decades, the prognosis for infants with Down’s syndrome has improved. From a population-based study conducted in the USA, the estimated median age at death among individuals with Down’s syndrome was 49 years in 1997 versus 25 years in 1983, an average increase in life expectancy of 1.7 years per year studied (Yang et al., 2002). Individuals with Down’s syndrome have more behavioural and psychiatric problems, including autism, than non-affected children (Roizen and Patterson, 2003). Institutionalization of infants with Down’s syndrome is now uncommon. Adults with Down’s syndrome have an increased risk of Alzheimer’s disease in their early fifties. Clinical signs and symptoms of Alzheimer’s disease are noted in 75% of adults with Down’s syndrome who are over 60 years of age (Roizen and Patterson, 2003).
The purpose of screening for Down’s syndrome is to identify affected pregnancies by an effective and safe method, thereby taking into account the woman’s concerns and preferences. The aim of the screening programme is to identify a subgroup at increased risk of Down’s syndrome and subsequently offer them confirmatory and invasive testing by either chorionic villus sampling or amniocentesis. These invasive procedures are associated with an increased risk of miscarriage of 0.8% following chorionic villus sampling and 0.3% following amniocentesis (Heckerling and Verp, 1991). The decision to proceed to invasive testing involves trade-offs of the benefits and risks.
Pregnant women may wish to undergo Down’s syndrome testing for reassurance that their unborn child does not have the disorder, to allow the option of termination if it does or to allow preparation for the birth of a child with the condition (Alfirevic and Neilson, 2004). The latter category includes planning delivery in a unit with adequate facilities to deal with the post-natal management of infants with Down’s syndrome. Some parents may seek adoption placement for their child with Down’s syndrome (Julian-Reynier et al., 1995).
Diagnostic tests
Initially, invasive testing such as chorionic villus sampling and amniocentesis was only offered to women of advanced aged (typically 35 years and over), but this approach identifies only 25–30% of fetuses with Down’s syndrome as the remainder of pregnancies complicated by fetal Down’s syndrome occur among younger women (Simpson, 2005). Chorionic villus sampling is typically performed between 10 and 14 weeks of gestation and amniocentesis at 15–18 weeks of gestation. Either procedure yields fetal cells from which chromosomal abnormalities may be identified.
Screening strategies
A glossary of the specific screening strategies being discussed in this section is summarized in Table II.
|
I. Non-invasive tests in the second trimester of pregnancy.
Non-invasive screening started with the observation that maternal serum concentration of 
-foetoprotein (AFP), used to screen for neural tube defects in the second trimester of pregnancy, tended to be lower in pregnancies complicated by Down’s syndrome (Cuckle et al., 1984). Later, other maternal serum analytes, including an elevated serum level of total hCG and a low serum level of unconjugated estriol (uE3) emerged as biochemical markers of Down’s syndrome in the second trimester of pregnancy, i.e. from 15 to 20 weeks’ gestation.
To allow for systematic changes in marker serum levels with increasing gestational age, serum concentrations are converted into multiples of the normal median (MOM) at a given gestational age for both affected and non-affected pregnancies. By using the observed estimates of likelihood ratios derived from maternal serum levels of the analytes, the individual risk of Down’s syndrome can be calculated using a mathematical model taking into account the maternal age-related a priori risk of the woman.
Women with a screening-derived risk of Down’s syndrome of 1/300–1/250 are typically considered to be at increased risk. The latter risk threshold points are roughly equivalent to the risk of a 35- and 36-year-old woman giving birth to a live born infant with Down’s syndrome (Table I). Invasive testing is offered to women whose test result indicates an increased risk of Down’s syndrome.
The use of ultrasonography to estimate gestational age improves the sensitivity and specificity of maternal serum screening (Benn et al., 1997). Other factors that are considered in non-invasive second-trimester screening include maternal weight, insulin-dependent diabetes mellitus, multiple pregnancy, ethnic origin, previous Down’s syndrome pregnancy and whether the test is the first one in a pregnancy or a repeat (Wald et al., 1997).
When an ultrasound scan is used to estimate gestational age, the detection rate of Down’s syndrome for a 5% false-positive rate is estimated to be 59% using the double test (AFP and hCG) and 69% using the triple test (AFP, hCG and uE3) (Wald et al., 1997). This approach requires approximately 60–70 amniocenteses to identify one fetus with Down’s syndrome. Second-trimester ultrasonographic findings, such as thickened nuchal fold of the fetus, can be used to adjust the screening-derived risk of trisomy 21 and, therefore, the need for amniocentesis (Smith-Bindman et al., 2001). Inhibin A, an alpha-beta subunit hormone of placental origin, is the latest addition to second-trimester serum screening. With the so-called quadruple test (AFP, hCG, uE3 and inhibin A), detection rates of 76–83% at a 5% false-positive rate can be achieved (Wald et al., 2004; Canick and MacRae, 2005; Malone et al., 2005).
II. Non-invasive tests in the first and second trimesters of pregnancy.
The efficacy of the various screening strategies was calculated in the Serum, Urine and Ultrasound Screening Study (SURUSS), a multicentre prospective study of 47 053 women with singleton pregnancies, including 101 pregnant women with Down’s syndrome (Wald et al., 2003; Wald et al., 2004) (Figure 1). The integrated test, comprising ultrasound measurement of the nuchal translucency (NT) of the fetus (Table II) and an assay of serum pregnancy-associated plasma protein-A (PAPP-A) measurements in the first trimester, combined with the serum levels of biochemical markers of the quadruple test in the second trimester (from 15 weeks), has the best screening performance, in terms of sensitivity and specificity (i.e. 86 and 94% for 1 and 5% false-positive rates, respectively) (Figure 1). The corresponding odds of being affected given a positive test result derived from the integrated test are 1:19 and 1:25, respectively. The serum-integrated test is a variant of the integrated test using serum markers only (PAPP-A in the first trimester and the quadruple test in the second trimester). The serum-integrated has a good screening performance as well (Figure 1). Concurrent with SURUSS in the UK, the First- and Second-Trimester Evaluation of Risk (FASTER) trial was conducted in the USA with the goal of providing direct comparative data on currently available screening approaches to Down’s syndrome (Malone et al., 2005). The FASTER trial involved 38 167 pregnant women, 117 of whom had a Down’s syndrome fetus. The findings of FASTER trial were comparable with those of SURUSS. At a 5% false-positive rate, the detection rate with serum-integrated screening was 88 and 96% with integrated screening (Malone et al., 2005).
|
The major disadvantage of the latter approaches is the delay in obtaining test results because women will not be informed of their results until the second trimester. This approach also precludes chorionic villus sampling for definite diagnosis and early termination of affected pregnancies if requested (Simpson, 2005). For this reason, women may prefer a screening approach that yields a test result in early pregnancy. Here, women could be offered stepwise sequential screening, in which they undergo first-trimester combined screening with the results provided immediately. Women whose first-trimester test results are indicative of an increased risk of Down’s syndrome (e.g. risk greater than 1:150) are offered chorionic villus sampling. Women with negative test results may chose to return at 15 weeks so that the quadruple markers can be measured, and a new risk estimate is given that combines the results of measurement of the first-trimester and the second-trimester markers (Malone et al., 2005).
III. Non-invasive tests in the first trimester of pregnancy.
The combination of ultrasound measurement of NT with maternal serum free ß-hCG and PAPP-A is currently considered the most effective first-trimester screening strategy. Depending on the gestational age and the maternal age distribution, this approach has an overall detection rate for Down’s syndrome of 80–91% at a false-positive rate of 5% (Crossley et al., 2002; Nicolaides, 2004; Malone et al., 2005; Cuckle, 2006; Perni et al., 2006). Using this approach, the estimated odds of being affected given a positive test result derived from the combined test ranges from 1:12 (Nicolaides, 2004) to 1: 19 (Malone et al., 2005) and 1: 27 (Wald et al., 2004).
Test performance in relation to maternal age.
Test performance, which refers to detection and false-positive rates derived from non-invasive screening test for Down’s syndrome, varies according to the age of the pregnant woman as is illustrated by the test performance of the combined test (Table III and Figure 2). The estimated probability that Down’s syndrome will be detected in a 20-year-old woman if she has a Down’s syndrome pregnancy is 63.5% with the combined test and 73.5% with the integrated test. The corresponding chance in a 40-year-old woman is 94.6 and 95.7%, respectively. The likelihood that the woman having a non-invasive test will be classified as having an ‘increased risk’ for Down’s syndrome increases with her age-related a priori risk (Table III and Figure 2). For instance, the estimated probability for a 20-year-old women being classified as having an ‘increased risk’ is 2.8% with the combined test and 2.4% with the integrated test, whereas the corresponding probabilities for a 40-year-old woman are 26.7 and 20.0%, respectively (Table III). All these risks may then be altered by the findings of a second-trimester genetic ultrasound when markers of fetal aneuploidy are sought.
|
|
What does an abnormal test result mean?
If the screening test is positive, indicating an increased risk of Down’s syndrome, invasive testing (chorionic villus sampling or amniocentesis) should be offered as an option rather than a self-evident act (Santalahti et al., 1998). The health care professional should provide the information and support necessary for the woman to make an informed choice about further testing. Informed choices are those based on relevant information that reflects women’s values (Marteau, 1995; Marteau and Dormandy, 2001).
The information provided should include the clarification of the screening-derived risk estimate for Down’s syndrome and other chromosomal abnormalities, where indicated, and the potential implications of confirmatory testing. In the intermediate-risk category during the first trimester, i.e. with a risk estimate of between 1 in 101 and 1 in 1000, further assessment of risk by detailed first-trimester ultrasound examination to determine presence/absence of the nasal bone, presence/absence of tricuspid regurgitation or normal/abnormal Doppler velocity waveform in the ductus venosus could increase specificity (Nicolaides et al., 2005; Avgidou et al., 2005; Spencer, 2006). Such ultrasound examination, however, requires sophisticated ultrasound equipment and a high level of expertise.
Low PAPP-A levels (<5th percentile), low levels of free ß-hCG (<1st percentile) and increased NT measurements (>99th percentile) in the first trimester of pregnancy are strongly associated with a number of adverse outcomes including spontaneous fetal loss, pre-eclampsia and fetal growth restriction (Dugoff et al., 2004; Krantz et al., 2004). The sensitivity of these findings, however, is quite poor (Spencer, 2006). The risk of fetal malformations, in particular of the heart and great arteries, is increased in euploid fetuses with an increased NT measurement (3.5 mm; >99th percentile) (Haak and van Vugt, 2003; Souka et al., 2005; Bilardo et al., 2006). The rise in risk is exponential as the NT measurement increases. Thus, an increased NT measurement is an indication for specialist ultrasound investigation, with emphasis on detailed ultrasound examination of the fetal heart and great vessels. If subsequent mid-trimester specialist ultrasound investigation is normal, a favourable outcome of pregnancy can be expected (Bilardo et al., 2006).
What does a normal test result mean?
If the test result is normal, the likelihood that the pregnancy is affected by Down’s syndrome is very small. The negative predictive value is close to 100% (Table III).
Conclusions and recommendations
Down’s syndrome screening is acceptable to the general public as clearly shown by the high acceptance rate of this ‘opt in’ screening service. Nevertheless, local audit data revealed large variations in uptake rates (Reynolds, 2003; Rowe et al., 2004; Dormandy et al., 2005; Van den Berg et al., 2005a; Müller et al., 2006). Low uptake rates of prenatal screening may be the result of limited access to prenatal testing, negative attitudes towards screening or both (Rowe et al., 2004). From a prospective study that was conducted in two hospitals in the UK, it was concluded that the relatively low uptake of screening for Down’s syndrome in women from minority ethnic groups and socioeconomically disadvantaged women does not reflect more negative attitudes towards screening but rather lower rates of informed decision-making (Dormandy et al., 2005). Informed choice is more likely to be associated with more realistic expectations of screening, with corresponding lower levels of emotional distress, and more satisfaction with the decision to participate or not in the screening programme (Marteau, 1995; Van den Berg et al., 2005b). A non-directive approach should be used and special attention paid to the notion that participation in a screening programme for Down’s syndrome is voluntary. Truly informed choice for participation or non-participation in a screening programme for Down’s syndrome may be difficult to achieve in practice. In this context, high acceptance rates may be misleading (Dormandy et al., 2002; Edwards et al., 2003).
Screening strategies that combine NT measurement with serum biochemical marker testing perform better than either of these tests alone. The integrated test is the most effective, safest, but most expensive approach (Gilbert et al., 2001). The choice of screening strategy should be between the integrated test, the serum-integrated, stepwise sequential test and the first-trimester combined test. Many women express a clear preference for first-trimester screening tests for Down’s syndrome (Mulvey and Wallace, 2000; Simpson, 2005). The second-trimester quadruple test, however, remains a relevant option for prenatal screening for Down’s syndrome because a considerable number of women do not seek prenatal care until the early second trimester. In fact, the introduction of second-trimester screening tests for Down’s syndrome has reportedly led to a sharp decrease (67%) of amniocenteses in non-affected pregnancies, in particular among older women (Benn et al., 2005). Interestingly, for non-invasive screening tests, the test performance, in terms of age-specific detection rates, increases with maternal age (Figure 2). A screening programme for Down’s syndrome based solely on maternal age is less effective, less safe and more costly than the above options (Gilbert et al., 2001).
|
Breast cancer |
|---|
|
TOPAbstractIntroductionPrenatal screeningBreast cancerOsteoporosisConclusionsReferences |
|---|
Condition of interest
Breast cancer is the most common cancer in women. The highest incidence (100–125/100 000 women) is found in the USA and Western Europe and the lowest (13–23/100 000 women) in China, South Central Asia and Africa. In Western populations, the average lifetime risk of breast cancer is about 5–10%. In 2005, there were 211 240 new cases of breast cancer among women in the USA. This accounts for 32% of all annual incident cases of cancer. More than 73 000 die of breast cancer annually, accounting for approximately 15% of cancer deaths among women in the USA (Jemal et al., 2005).
The prevalence of several established risk factors differs across racial and ethnic subpopulations and may contribute to the higher incidence rates in White women compared with other racial and ethnic groups (Ghafoor et al., 2003; Ward et al., 2004).
Overall, the increase in female breast cancer incidence may be attributable to increased use of hormone replacement therapy and delayed childbearing (Nelson et al., 2002; Beral and Million Women Study Collaborators, 2003; Ghafoor et al., 2003; Minelli et al., 2004; Collins et al., 2005; Greiser et al., 2005; Jemal et al., 2005). The increased incidence of breast cancer may also reflect the increased use of screening by mammography (Ghafoor et al., 2003). More than 50% of all breast cancers are diagnosed in women aged 65 years, and almost 75% are in post-menopausal women (Fracheboud et al., 2004).
Women with one or more affected first-degree relatives are at increased risk. The magnitude of risk depends on the number of affected first-degree relatives and the presence of a pathogenic mutation in a breast cancer associated gene, such as BRCA1 and BRCA2, among others (Antoniou et al., 2003; Meijers-Heyboer, 2006). The US National Institutes of Health made available a computerized tool for calculating the individual woman’s risk of breast cancer (http://www.cancer.gov/bcrisktool/). This risk assessment tool, however, does not take into account cancer risks associated with mutations in breast cancer-associated genes. Women with mutations of the BRCA1 and BRCA2 genes have considerably higher lifetime risk (up to 85%) of breast cancer than the unselected general population. They are also at increased risk of ovarian cancer. However, mutations in these genes are rare in the general population and account for only a small fraction of all breast cancer cases and for less than one-fifth of the familial risk of breast cancer (Anonymous, 2000). In Western populations, the estimated combined prevalence of BRCA1 and BRCA2 mutations is 0.2% in the general population, 6% in women diagnosed with breast cancer before the age of 50 years and 1.3% among women with breast cancer who are 50 years of age at the time of diagnosis (Peto et al., 1999).
The aim of screening is the early detection and treatment of women with breast cancer and ultimately to reduce morbidity and particularly mortality from this condition. The notion of early detection of breast cancer relates to the aim of detecting non-invasive stages of breast cancer [ductal carcinoma in situ (DCIS)] or early invasive breast cancer.
There are several screening tests available for the early detection of breast cancer.
Breast self-examination
Breast self-examination (BSE) is an intuitively attractive concept, because theoretically, a well-trained woman who practices BSE might improve her survival by detecting breast masses when they are relatively small. However, palpable breast masses are common and usually benign, particularly in young women. BSE may lead to unwarranted anxiety, false reassurance and unnecessary medical interventions (Baxter and the Canadian Task Force on Preventive Health Care, 2006). From the literature to date, BSE has not been shown to be effective in reducing breast cancer mortality (Baxter and the Canadian Task Force on Preventive Health Care (2006); Elmore et al., 2005).
Clinical breast examination
Few data about the efficacy of clinical breast examinations are available from randomized clinical trials. Four randomized trials of mammography included the clinical breast examination in the screened group (Barton et al., 1999; Elmore et al., 2005). On the basis of a meta-analysis, the estimated sensitivity of clinical breast examination was 54% (95% CI: 48–60%) and specificity 94% (95% CI: 90–97%). There are no randomized trials comparing clinical breast examination with a control group that received no screening.
Mammography
Mammography is the best tool available for screening for breast cancer (Elmore et al., 2005). The accuracy of regular mammography screening varies with age, with sensitivities ranging from 68–88% and specificities from 82–98.5% (Fletcher and Elmore, 2003). DCIS of the breast has become more common because it can present as microcalcifications detected by mammography. DCIS accounts for approximately 13–20% of breast cancers diagnosed by mammography (Leonard and Swain, 2004; Advisory Committee on Breast Cancer Screening, 2006).
Other imaging techniques
Magnetic resonance imaging (MRI) is time-consuming and expensive and, therefore, not an appropriate screening test in unselected populations (Fracheboud and de Koning, 2006). MRI may be of value in the screening of high-risk women. Sensitivity of MRI in high-risk women has been found to be much higher than that of mammography, but specificity is generally lower (Elmore et al., 2005). Ultrasonography can detect cystic masses, which are common, and may be used to guide biopsy techniques. Mammoscintigraphy and positron emission tomography (PET) may be helpful as an adjunct to clinical examination and mammography. They have not yet been tested as a screening instrument in larger populations.
Genetic testing
There are several mutations at one of more genetic loci involved in families with familial breast cancer, most notably BRCA1 and BRCA2.
What does an abnormal test result mean?
A thorough clinical breast examination, imaging and tissue sampling are needed to identify malignancy. Where indicated, tissue specimens obtained with fine-needle biopsy allow histological diagnosis, hormone-receptor testing and differentiation between in situ and invasive diseases. Large differences have been noted between the percentage of screening mammograms considered abnormal (known as the ‘recall rates’) within community-based mammography programmes in the USA and those in other countries. For example, the recall rate in the USA (about 13%) is considerably higher than that in the UK (about 8%), with no difference in cancer detection rate per woman screened (Smith-Bindman et al., 2003). Overall, about nine out of 10 women with abnormalities on the mammogram do not have breast cancer (Elmore et al., 2005; Fracheboud and de Koning, 2006). Nonetheless, the likelihood of having breast cancer when the mammogram is abnormal depends heavily on the woman’s age and clinical findings.
There are various treatment regimens available for breast cancer:
Surgery.
The mainstay of care for patients with early breast cancer is surgical therapy. Depending on the stage of breast cancer at diagnosis and other risk factors, surgery may be breast-conserving (lumpectomy) or mastectomy, with or without axillary clearance. Surgery may be preceded by chemotherapy. Radiotherapy with or without adjuvant systemic therapy may be applied after surgery. Controversy exists with regard to the optimal management of women with DCIS (Julien et al., 2000; Westenberg et al., 2003; Baxter et al., 2004; Leonard and Swain, 2004). Lumpectomy with or without radiation therapy is currently considered the standard treatment option for DCIS.
Chemotherapy
From a Cochrane review, it was concluded that several months of polychemotherapy is typically associated with highly significant reductions in recurrence risk [OR: 0.76 (95% CI: 0.73–0.80)] and mortality risk (from all causes) [OR: 0.85 (95% CI: 0.80–0.81)] (Early Breast Cancer Trialists’ Collaborative Group, 2002). The age-specific benefits of polychemotherapy appear largely irrespective of menopausal status at presentation, of estrogen receptor (ER) status of the primary tumour and of whether adjuvant tamoxifen had been given.
Radiotherapy.
On the basis of information available on 42 000 women in 78 randomized treatment comparisons, Clarke et al. (2005) concluded that radiotherapy regimes are associated with a 20% relative reduction of the risk of local recurrences among women who have a substantial recurrence risk (>10%). Among this latter group of women, a 5% relative reduction of 15-year breast cancer mortality was also observed (Clarke et al., 2005). There was, however, a statistically significant excess of non-breast-cancer mortality in irradiated women. The excess mortality was mainly from heart disease (rate ratio 1.27) and lung cancer (rate ratio 1.78) (Clarke et al., 2005; Darby et al., 2005). Adjuvant radiotherapy does not seem to confer any benefit in women who already have a low local recurrence risk (<10%).
Tamoxifen and aromatase inhibitors.
In women with early breast cancer who have ER-positive tumours, 5 years of adjuvant tamoxifen treatment, an anti-estrogen that is primarily active against the effects of estrogens on breast tissue, reduces the annual breast cancer death rate by 31% (Early Breast Cancer Trialists’ Collaborative Group, 2005). There is insufficient evidence of a beneficial effect of tamoxifen in women with receptor-negative breast cancer after chemotherapy (Early Breast Cancer Trialists’ Collaborative Group, 2001). Aromatase inhibitors, including the non-steroidal agents anastrozole and letrozole and the steroidal agent exemestane, are only effective in post-menopausal women and can be used as extended adjuvant therapy after 5 years of tamoxifen treatment (Thurlimann et al., 2005; Smith and Chua, 2006).
Oophorectomy.
In premenopausal women with early breast cancer, ablation of functioning ovaries significantly improves long-term survival (Early Breast Cancer Trialists’ Collaborative Group, 2000). Oophorectomy causes immediate and often severe menopausal symptoms, carries an increased risk of osteoporosis and is inevitably associated with sterility (Smith and Chua, 2006). Oophorectomy has now been largely abandoned in premenopausal women in favour of tamoxifen treatment.
Treatment for those positive on genetic testing.
In women with a family history of breast cancer and who test positive for a BRCA1 or BRCA2 mutation, decisions as to whether to undergo prophylactic risk-reduction surgery vary—some will choose prophylactic bilateral mastectomy and some will opt for intensive surveillance (Nelson et al., 2005). Because both subcutaneous and total mastectomies result in incomplete removal of all breast tissue, it is important that women are aware that breast cancer can still occur after prophylactic bilateral mastectomy (Eisen et al., 2000). A corollary of such genetic testing is that some women who are positive for one or other of these mutations may subsequently request prenatal diagnostic testing with the option of pregnancy termination if the fetus is affected.
What does a normal test result mean?
When on clinical examination the breasts appear normal and the mammogram is considered as negative (assessment category 1 or 2), the woman can be reassured as her odds of having the disease are 10 times lower than they were before having the test (Elmore et al., 2005).
Conclusions and recommendations
Screening mammography reduces breast cancer mortality by about 25–35% in women aged 50–69 years at more than 10 years of follow-up (Humphrey et al., 2002; Elmore et al., 2005; Advisory Committee on Breast Cancer Screening, 2006; Fracheboud and de Koning, 2006). Compared with older women, the absolute benefit of screening women in their forties is smaller (i.e. approximately 11–18%) because of a lower incidence of disease, denser breast tissue (which lowers test performance, in terms of sensitivity) and on average faster growing tumours (Elmore et al., 2005; Moss et al., 2005). From a report by the Advisory Committee on Breast Cancer Screening (2006), it was concluded that the NHS breast screening programme in the UK reduces mortality from breast cancer at the cost of only £3000 per year of life saved. The estimated number of women needed to screen regularly (i.e. at 3-year intervals) over a 10-year period to save one life is 400 (Advisory Committee on Breast Cancer Screening, 2006; Dixon, 2006). Although screening mammography is associated with a meaningful reduction of breast cancer mortality, there is some controversy about the impact of screening on long-term overall survival (Olsen and Gøtzsche, 2001; Tabár et al., 2002; de Koning, 2003; Freedman et al., 2004). There are also concerns about overdiagnosis and subsequent unnecessary treatment of clinically insignificant disease (Gøtzsche, 2004). Although there is an element of overdiagnosis of DCIS in breast cancer screening, the phenomenon is small in both relative and absolute terms (Yen et al., 2003; Duffy et al., 2005; Advisory Committee on Breast Cancer Screening, 2006). There is a delicate balance between reducing recall rates, so far that small cancers are not missed, and calling back too many women, which causes anxiety and may reduce reattendance (Advisory Committee on Breast Cancer Screening, 2006). Other limitations of screening mammography include inconvenience, pain and monetary cost. Women with screen-detected disease may experience psychological effects of living with the diagnosis for longer than would otherwise be the case (Elmore et al., 2005; Advisory Committee on Breast Cancer Screening, 2006). This emphasizes the importance of providing comprehensive information about screening including the tests, treatment options and likely prognoses to enable an individual woman to make a fully informed choice about participation in the programme (Fletcher and Elmore, 2003; Thornton et al., 2003).
Genetic testing for BRCA1 and BRCA2 mutations is laborious and expensive, making careful case selection desirable. Many clinics use a 10% probability threshold to identify high-risk women for targeted screening (Meijers-Heyboer, 2006).
|
Osteoporosis |
|---|
|
TOPAbstractIntroductionPrenatal screeningBreast cancerOsteoporosisConclusionsReferences |
|---|
Condition of interest
Osteoporosis is a condition resulting in an increased risk of skeletal fractures due to a reduction in the density of bone tissue (Consensus Development Conference, 1991). It is characterized by low bone mass and disordered microarchitecture of bone tissue, which cause increased bone fragility and risk of fracture (Johnson, 2006). Vertebral fractures can be debilitating; they are associated with increased functional impairment, back pain and kyphosis (Van der Klift et al., 2002).
Several studies have noted that the presence of a vertebral fracture is associated with an increased mortality risk (Center et al., 1999; Ensrud et al., 2000). More important in terms of frequency is fracture-related morbidity. In the USA, approximately 10% of patients are disabled by hip fracture and 19% require institutionalization (Melton, 2003). The major problem in studying the occurrence of vertebral fractures is the lack of a consensus on definition (Melton, 2000). On the basis of a prospective population-based cohort study of men and women aged 55 years (the Rotterdam Study), the incidence of vertebral fractures correlates strongly with age: for women ranging from 7.8/1000 person-years at ages 55–65 years to 19.6/1000 person-years at ages >75 years (Van der Klift et al., 2002). Overall, the incidence of vertebral fractures is higher in women than in men.
In general, hip, wrist and upper humerus fractures are the most frequent non-vertebral fractures in both men and women (Schuit et al., 2004). Estrogen deficiency and changes in vitamin D metabolism are important contributors to the occurrence of osteoporosis in post-menopausal women. Another important factor in age-related bone loss is the decrease in calcium absorption that normally occurs (Gallagher et al., 1979). In the USA, the lifetime risk of hip fracture is at least 17.5% in White women (Melton, 2000). The cumulative lifetime fracture risk for a 50-year-old women may be as high as 60% (Cummings et al., 1989). Furthermore, subjects with vertebral fractures have an increased risk of both new vertebral and non-vertebral fractures such as hip fractures (Van der Klift et al., 2002). There has been increasing attention paid recently on the potential association of fracture with serum homocysteine levels (van Meurs et al., 2004).
Annual expenditures for osteoporotic fracture care in the USA ($17.5 million in 2002) are dominated by hip fracture treatment but vertebral fractures, distal forearm fractures and the other fractures related to osteoporosis contribute one-third of the total (Melton, 2003).
The goal of osteoporosis screening is to identify women at increased risk of fracture and subsequently to reduce that risk by the introduction of treatment intended to prevent further bone density loss.
A dual-energy X-ray absorptiometry (DEXA) scan is the current gold standard test for the diagnosis of osteoporosis. In post-menopausal women, the T-score for bone mineral density (BMD) is a well-accepted diagnostic criterion for osteoporosis. Although BMD can be measured by DEXA at a peripheral site (e.g. wrist or heel), central measurements such as those at the femoral neck or lumbar spine are the most useful (de Laet et al., 2002).
The sensitivity of using a T-score for femoral neck BMD at or below –2.5 for identifying both men and women at risk of non-vertebral fractures has been assessed prospectively in a population-based cohort study. Only 44% of all non-vertebral fractures occurred in post-menopausal women with a T-score below –2.5; in men, this percentage was even lower (21%) (Schuit et al., 2004).
What does an abnormal test result mean?
The incidence of vertebral fractures doubles per SD decrease in lumbar spine or femoral neck BMD (Van der Klift et al., 2002). For all non-vertebral fractures, the age-adjusted hazard ratio (95% CI) per SD decrease in femoral neck BMD is 1.5 (1.4–1.6) (Schuit et al., 2004). One systematic review noted that there is some evidence that exercise (such as aerobics, weight bearing and walking) is effective at 1 year or longer in slowing bone loss in post-menopausal women, although it had no effect on fracture risk (Bonaiuti et al., 2002). Although this may be the easiest and simplest measure to implement long-term, there is currently no evidence that such a strategy is effective in terms of health outcome. Exercise and hormone replacement therapy have been shown to improve bone density (Prince et al., 1991; Wells et al., 2002). Nonetheless, because of the associations of hormone-replacement therapy with both breast cancer and coronary heart disease, as was demonstrated in the Women’s Health Initiative randomized trial (Rossouw et al., 2002), the use of hormones among post-menopausal women has declined dramatically (Finkelstein, 2006). The effectiveness of calcium and vitamin D supplementation to prevent fractures is limited (Avenell et al., 2005; Jackson et al., 2006). Women receiving calcium plus vitamin D supplementation have a slightly increased risk of kidney stones (hazard ratio 1.17; 95% CI: 1.02–1.34) (Jackson et al., 2006). Alternative pharmacologic treatment options for post-menopausal women at increased risk of fractures include anti-resorptive drugs such as raloxifene (Cranney et al., 2002a), calcitonin (Cranney et al., 2002b) and the bisphosphonates [alendronate (Cranney et al., 2002c), risedronate (Cranney et al., 2002d) and ibandronate (Cooper et al., 2003; Felsenberg et al., 2005)]. These drugs improve bone density and reduce fracture risk but are of limited value in halting further deterioration of skeletal microarchitecture. The newest agent for the treatment of post-menopausal women at increased risk of fractures is recombinant human parathyroid hormone, teriparatide, which increases bone mass and also restores bone architecture and integrity (Neer et al., 2001; Body et al., 2002). Evidence is still required, however, for the effectiveness and cost-effectiveness of screening programmes in terms of the timely identification and subsequent treatment options for osteoporosis and/or increased fracture risk.
What does a normal test result mean?
A normal test result indicates that bone density is adequate and that the risk of fracture is not increased because of osteoporosis.
Conclusions and recommendations
Osteoporosis and its consequences, particularly vertebral and hip fractures, are serious public health problems for both older men and older women (Cummings and Melton, 2002). The average hip fracture risk in women is much higher than in men but appears to be similar at the same BMD (de Laet et al., 2002). Although bone densimetry by DEXA scan is a well-established tool for the diagnosis of osteoporosis, its unselective use in a screening setting is largely ineffective, and costly, because of the low sensitivity for bone fractures (Melton et al., 2004). There is a clear need for the development of more sensitive risk assessment tools, using not only bone densimetry but also other powerful predictors of fractures.
Clinical decision-making is currently limited to treating patients with fractures, who presumably have already failed any public health measures in place, or to patients with low bone density identified by case-finding (Melton et al., 2004). In fact, the presence of a vertebral fracture and a low BMD are both strong independent predictors of—recurrent—vertebral fractures (Van der Klift et al., 2002). The tools needed to predict the risk of an osteoporotic fracture over the next 10 years are now being developed (Melton et al., 2004). Apart from bone density measures, these include, for example, smoking, low weight, a history of osteoporotic fracture or hip fracture in first-degree relatives, menopause before the age of 45 years and glucocorticoid use (Kanis et al., 2004; Schuit et al., 2004; Kanis et al., 2005; Johnson, 2006) and/or laboratory parameters (Raisz, 2004). An increased circulating homocysteine level is a potentially reversible risk factor for osteoporosis. The overall relative risk of fracture was 1.4 (95% CI: 1.2–1.6) for each increase of 1 SD in the natural-log-transformed homocysteine level (van Meurs et al., 2004) after adjusting for confounding factors. Using data from the Framingham Study, McLean et al. (2004) found that the risk of hip fracture was increased by nearly a factor of two in women for the highest quartile of plasma homocysteine levels compared with the lowest quartile. The associations between homocysteine levels and the risk of fracture appeared to be independent of BMD and other potential risk factors for fracture. It is uncertain, however, whether there is a direct link between increased homocysteine levels and fractures (Raisz, 2004). Homocysteine is associated with cardiovascular disease and cognitive dysfunction, conditions that contribute to increased frailty and, in turn, tendency to fall.
|
Conclusions |
|---|
|
TOPAbstractIntroductionPrenatal screeningBreast cancerOsteoporosisConclusionsReferences |
|---|
The purpose of this review, and specifically the reason for selecting these three conditions, is that they illustrate women’s health issues where early detection and treatment can have considerable potential for improving both the quality and longevity of life. It is crucial that women are provided with clear and comprehensive information about the screening programme, both in terms of possible gains and in terms of costs of various kinds—physical, economic and psychological. Informed choice requires the full disclosure of test details (their nature and ability to detect or rule out the condition of interest) and the implications for treatment and prognosis. In contrast to other public health measures such as vaccination, the concept that screening programmes require full coverage of the relevant population in order to be worthwhile should be challenged if we are to achieve the right balance between the individual and the public health perspective. Further research is required into individuals’ valuations/utilities of the different health outcomes that might accrue from participation or non-participation in the programme (Heckerling and Verp, 1991
; Coast, 2004).
For health care providers involved in the care of women of reproductive age, it is important to discuss with them the possibilities and implications of prenatal screening. From the public health perspective, it is important to choose a screening strategy that is both safe and effective, at acceptable capital costs (Wilson and Jungner, 1968; Gilbert et al., 2001). Prenatal screening for Down’s syndrome should not be focused on cost avoidance or eradication of individuals with a disability or handicap. It is primarily concerned with providing couples with accurate information by which they can make reproductive choices (Spencer, 2006).
Regarding breast screening, the evidence is more persuasive for older women (50–69 years) than those under 50 years of age, but even in the former group, there is a balance of risks and benefits for individual women (Fletcher and Elmore, 2003; Thornton et al., 2003). In any event, the occurrence of cancers between screening tests means that screening is not a panacea: there is no guarantee for women that their breast cancer will be detected if they participate in the programme.
Treatment options for osteoporosis have improved in recent years in terms of reductions in fracture risk as well as beneficial effects on bone density, but there is currently a lack of evidence regarding screening programmes. Nonetheless, it is very unlikely that screening involving unselective bone scans of the population would be cost-effective. Evidence is still needed regarding the feasibility, acceptability and cost-effectiveness of more targeted approaches incorporating risk factors. This will still require randomized evaluations of the putative screening programme within pragmatic trials that have health gain as the primary outcome.
Because screening is in essence a sophisticated form of risk assessment, a major challenge in operationalizing all such screening programmes is the communication of risks, which requires improved understanding of risk perception (Edwards et al., 2003; Barratt et al., 2004; Rimer et al., 2004). Such information should be presented in imaginative ways including the classical performance measures of screening tests such as detection rates, proportions requiring further diagnostic workup and false-positive risks, along with broader issues such as the number of women needed to screen in order to detect one individual with the condition of interest and the number of women needed to screen in order to avoid one poor outcome. We still have a long way to go.
|
Acknowledgements |
|---|
|
TOPAbstractIntroductionPrenatal screeningBreast |
|---|